Sequential Aggregation and Rematerialization: Distributed Full-batch Training of Graph Neural Networks on Large Graphs.
Hesham MostafaPublished in: CoRR (2021)
Keyphrases
- neural network
- graph representation
- graph theory
- training process
- weighted graph
- directed graph
- adjacency matrix
- graph structure
- graph matching
- graph construction
- graph theoretic
- graph mining
- labeled graphs
- distributed sensor networks
- graph databases
- graph partitioning
- graph model
- bipartite graph
- training algorithm
- graph structures
- graph theoretical
- subgraph isomorphism
- graph properties
- random graphs
- graph clustering
- graph search
- graph data
- distributed systems
- series parallel
- dynamic graph
- graph isomorphism
- pattern recognition
- graph classification
- undirected graph
- structural pattern recognition
- graph representations
- batch mode
- feedforward neural networks
- graph kernels
- multi layer perceptron
- community discovery
- training set
- graph transformation
- feed forward neural networks
- dense subgraphs
- spanning tree
- directed acyclic
- query graph
- topological information
- small world
- evolving graphs
- connected graphs
- neighborhood graph
- graph layout
- maximum cardinality
- real world graphs
- minimum spanning tree
- random walk
- maximum clique
- edge weights
- graphical structure
- frequent subgraphs
- back propagation
- planar graphs
- structured data
- neural network model
- connected dominating set
- inexact graph matching
- finding the shortest path
- bounded treewidth
- connected components
- reachability queries
- artificial neural networks
- graphical models
- shortest path
- adjacency graph